WO2016124066A1

WO2016124066A1 - Image verification code generation method and system

Info

Publication number: WO2016124066A1
Application number: PCT/CN2016/070553
Authority: WO
Inventors: 赵志猛; 罗治凡; 闪光辉; 刘深
Original assignee: 北京京东尚科信息技术有限公司; 北京京东世纪贸易有限公司
Priority date: 2015-02-05
Filing date: 2016-01-11
Publication date: 2016-08-11
Also published as: HK1208951A1; CN104700441B; CN104700441A

Abstract

An image verification code generation method and system, the method comprising: randomly selecting, from a character set, a set of character strings comprising a plurality of characters, lengths of the character strings being randomly determined (S101); randomly selecting a font from a font set for each character in the character strings, and rendering to obtain rendered characters, the fonts in a font library being a set of size n × m comprising n fonts and m word sizes, and n and m being preset natural numbers (S102); distorting each of the rendered characters to obtain distorted characters (S103); generating an image comprising a plurality of the distorted characters (S104); and generating at least one interference line on the image to obtain an image verification code, a path of the interference line being randomly generated (S105). The method improves user experience of the verification code, increases the difficulty of recognizing the verification code by a computer, and effectively ensures system security.

Description

Image verification code generation method and system

Technical field

The present invention relates to the field of verification code related technologies, and in particular, to a picture verification code generation method and system.

Background technique

Because of the ease of use and other reasons, the mainstream verification code in the Internet is now the picture verification code. Image verification code is closely related to image processing technology, such as embedding of characters, addition of noise, etc., and basic image processing knowledge is required. The cracking of the image verification code is more complicated. It requires not only professional image processing knowledge, but also knowledge of artificial intelligence. However, with the development of artificial intelligence and the popularity of open source tools, simple image verification code recognition It is getting easier and easier.

At present, the design of image verification code can be roughly divided into the following three types:

(1) The most basic and most popular use is a verification code with random characters and numbers as picture content. The length of this verification code character is usually fixed (length is 4). In the case of random characters, corresponding interference elements such as background color, interference line, etc. can be added as needed. In this verification code system, according to the characteristics of the interference elements, it can be divided into two types: 1. A relatively fixed interference factor is adopted. This kind of verification code interference element looks quite big to the naked eye, but it doesn't change much to the computer, such as the background color. Even if the background color changes constantly, it is usually very easy to remove; for example, the position of the character or even the font It is fixed or the whole character of the string drifts left and right. In fact, it is easy for the computer to find the area where the character is located. This kind of verification code is easier to crack. 2. Interference factors have a certain randomness. This kind of verification code interference factor has a large randomness, such as the position of the interference line, the font of the character, the relative position of the character, the rotation angle of the character, etc. are constantly changing, but the user experience of this verification code is generally not Too good, such as character rotation, usually the character selection direction and angle are not fixed. When the naked eye recognizes, people need to subconsciously tilt the head left and right to see clearly; the change of character font is usually taken in different pictures. Different font styles, and most of the fonts in the same picture are the same, even if using a variety of fonts, such as: hollow fonts, solid fonts, upper and lower color processing, shadow processing, etc., in fact, these are nothing more than The verification code is merged into one piece, and the difference between the two pictures is too large, which affects the user experience.

(2) Use professional knowledge as the verification code for the picture content. For example: a relatively strong professional mathematics website can use the points commonly used in mathematics as the content of the verification code, and the verification code needs the result of the integration; the website with strong professionalism can use a similar "current unit is ohm". You only need to judge Correctness will do.

(3) Use complex images as the content of the verification code. For example, a complete picture with a pattern, the middle of a circular area to do a certain rotation, the verification needs to rotate the middle area to restore the pattern into a complete image to verify correct; or a small with a few small Dog pattern, you need to answer the puppy in the pattern.

Disadvantages of prior art solutions

The existing seemingly complicated interference factors actually have very little interference on the computer, and some seemingly simple interference factors are difficult to identify. For the traditional picture verification code system, the effect of the background color is almost negligible. In order to be better recognized by the naked eye, the color of the character must be clearly distinguished from the background color, if it is not well distinguished. The difficulty of human visual recognition will greatly affect the user experience, but if the background color is easy to distinguish, then those seemingly complicated background colors are very easy for the computer, so the background color will not function as it should, and Complex background colors have serious interference effects on some weak and color-blind users, which is not conducive to their use; some verification codes use character distortion as a disturbance factor, and existing distortions are performed by character rotation, because each character The orientation of the options is different, so it will be more difficult to identify with the naked eye;

For some complex verification codes, such as each character is distorted, and the distortion strength and direction of each character are different, this verification code machine is difficult to identify, but the recognition of human eyes is very difficult, which is not conducive to users. Experience, these are obviously not conducive to the user experience, After all, the network system is to serve people. The first role of the system should be to provide services, not to prevent others from illegally using it.

Summary of the invention

Based on this, it is necessary to provide a method and system for generating a picture verification code, in view of the technical problem that the prior art fails to provide a picture verification code that is more easily recognized by the naked eye but is more difficult to be solved by the computer.

A picture verification code generation method includes:

The string selection step includes: randomly selecting, from the character set, a set of characters including a plurality of characters, the length of the string being randomly determined;

The font rendering step includes: randomly selecting a font and a font size from a font set for each character included in the string to perform rendering processing, and the font library includes a plurality of fonts, and each font has multiple types. Font size, the n and m are preset natural numbers;

The font warping step includes: twisting each of the rendered characters to obtain a twisted character;

a picture pre-generation step, comprising: generating a picture including a plurality of the twisted characters;

The interference line adding step includes: generating at least one interference line on the picture to obtain a picture verification code, and the trajectory of the interference line is randomly generated.

A picture verification code generation system includes:

a string selection module, configured to: randomly select a group of characters including a plurality of characters from a character set, and the length of the string is randomly determined;

a font rendering module, configured to: randomly select a font and a font size from a font set to render characters for each character included in the character string, where the font library includes multiple fonts, and each font has multiple fonts. Kind size

a font warping module, configured to: perform distortion processing on each of the rendered characters to obtain a twisted character;

a picture pre-generation module, configured to: generate a picture including a plurality of the twisted characters;

The interference line adding module is configured to: generate at least one interference line on the picture to obtain a picture verification code, and the trajectory of the interference line is randomly generated.

The interference element of the present invention uses only the interference line easily recognized by the naked eye, and the distortion of the character instead of the rotation of the character. These factors greatly improve the user experience of the verification code; while the length of the character is not fixed, the interference line trajectory The unfixed characters and different fonts increase the difficulty of the computer identification verification code, which effectively ensures the security of the system.

DRAWINGS

1 is a flowchart of a method for generating a picture verification code according to the present invention;

Figure 2 is a flow chart showing the operation of a preferred embodiment of the present invention;

FIG. 3 is a structural block diagram of a picture verification code generating system according to the present invention.

detailed description

The present invention will be further described in detail below in conjunction with the drawings and specific embodiments.

FIG. 1 is a flowchart of a method for generating a picture verification code according to the present invention, including:

Step S101, comprising: randomly selecting, from the character set, a set of character strings including a plurality of characters, where the length of the character string is randomly determined;

Step S102 includes: randomly selecting a font and a font size from a font set for each character included in the character string, and performing rendering processing to obtain a rendered character, where the font library includes multiple fonts, and each font has multiple font sizes. ;

Step S103, comprising: performing distortion processing on each of the rendered characters to obtain a twisted character;

Step S104, comprising: generating a picture including a plurality of the twisted characters;

Step S105, comprising: generating at least one interference line on the picture to obtain a picture verification code, and the trajectory of the interference line is randomly generated.

Wherein, in step S101, the character set is preferably a filtered number, a group of letters A string of characters that removes confusing characters in the picture, such as I and 1, 6, and b.

In step S102, the font set includes n types of fonts, and each font has m font sizes, for example, a small three-character font, a small three-character Song, a small three-character, a third-sized font, and a third font. The body of the Song, the size of the third size, the body of the small four-character, the Song of the small four-character, and the circle of the small four-character. For each string, for example: 4B69, the characters included are: 4, B, 6, and 9. Select a font and a font size for each character, for example, 4 selects the small three-sized body, and B selects the third size. The round body, 6 selects the small quadrilateral circle, and 9 selects the 3rd font size. The font size and font of each character are different. For different font sizes and different font characters, the human eye will be easily recognized, and the machine is It is necessary to extract features for recognition. When the fonts selected by different font sizes are different, it is necessary to extract features for different fonts with different font sizes, and the fonts and font numbers selected by different characters are random, and the difficulty of machine recognition is greatly increased.

The distortion of the character in step S103 is only a distortion of the shape of the character, instead of rotating the character, since the distortion of the shape of the character is only possible, the human eye can still recognize, but for the machine, since it does not know What kind of distortion is made in the character shape, its characteristics are not available, and the difficulty is greatly improved.

The interference line added in step S104 can be well distinguished from the verification code for the naked eye, but for the machine, the added interference line will cause the recognition of the character to be unable to extract suitable features, resulting in The difficulty of recognition increases.

In one of the embodiments:

And a verification code parameter obtaining step, the verification code parameter obtaining step includes: obtaining, according to the verification code providing request, the application identifier included in the verification code providing request, and acquiring the verification code parameter corresponding to the application identifier, The verification code parameters include a minimum allowable length and a maximum allowable length of the character string, a picture color, and a picture resolution;

And in the string selection step, the length of the character string is randomly determined between the minimum allowable length and the maximum allowable length;

And in the picture pre-generation step, the picture is determined according to a picture color and a picture resolution included in the verification code parameter.

This embodiment enhances the versatility of the verification code system, because a set of verification codes will be applied to multiple systems in actual operation, but the characteristic parameters sometimes required by different systems for better user experience machine security needs. Different, for example, the size of the image required by the mobile terminal will vary according to the resolution. The color of the image will change according to the theme color, so an easy-to-operate configuration system is required for personalization of different applications. Therefore, the verification code system can be applied to multiple applications, and the configuration of the parameters can be completely changed at random on the line, which enhances the flexibility of the system wind control.

In this embodiment, different verification code parameters may be set in advance for different applications, thereby controlling the length of the character string, the color of the picture, and the resolution of the picture.

In one embodiment, the twisting process includes: selecting at least one portion on each of the rendered characters as a portion to be twisted, and performing an upward translation, a downward translation, a left translation, or Pan to the right.

In this embodiment, the distortion of the character string is not a rotation of the character but a translation of the upper and lower sides of the different parts of the character, preferably, the translation is a smooth translation. Since only the different parts of the character are translated, it is still well recognized by the naked eye, but for the machine, since the machine generally performs feature recognition, that is, the color that is distinct from the background color on the picture, A plurality of different points are extracted as feature points thereon, and then the same or similar words of the feature points are selected from the database according to the location of the feature points. However, with the twisting method of the present embodiment, the position of the feature point of the verification code has been changed, and the changed feature point is randomly selected, so that it is difficult for the machine to recognize it.

In one of the embodiments, each of the interference lines is generated as follows:

A plurality of track points are randomly selected on the picture to generate a plurality of equation tracks including all of the track points.

Preferably, the interference line determines a cubic equation trajectory using four random points in the picture. The interference line is added. Since the machine does not know what kind of font its body is, it can only extract feature points when it is recognized. Therefore, adding interference lines is equivalent to increasing the feature points when the machine recognizes, but the added feature points are It is not the actual feature point of the verification code, so it is difficult for the machine to recognize it.

The way to determine a multi-equation trajectory through multiple points can be implemented by the prior art, such as the tool class implementation in the java library.

In one of the embodiments, the interference line is the same color as the character, and the thickness of the interference line is gradually changed between a maximum width and a minimum width of all the twisted characters.

In this embodiment, the color of the interference line is the same as the color of the character, the machine is more difficult to recognize, and the thickness of the interference line is gradual, which further improves the recognition difficulty of the machine.

2 is a flow chart showing the operation of a preferred embodiment of the present invention, including:

Step S201, obtaining a request for providing a verification code for the user;

Step S202: Acquire corresponding verification code parameters from a configuration management system (CMS) according to the request information. The verification code parameters can be: the maximum length and minimum length of the character, the color of the picture, the effective time of the picture, the minimum interval between the two picture requests, and so on. The generation of the verification code is performed under the control of related parameters.

In step S203, a random string is generated. A character string of a corresponding length is randomly selected in the character set as a random character string, and the length of the random string is randomly selected between the maximum length and the minimum length of the character. This character set is a filtered string of numbers and letters, which removes characters that are confusing in the picture, such as I and 1, 6, and b.

Step S204, random string rendering.

The length and width of the image are read from the CMS to generate a background image of the corresponding size, and the random string is rendered on the background image. During the rendering process, each character of the random string is randomly selected from the font set to be initialized. The font in the font set is a collection of n×m size consisting of n fonts and m font sizes; After rendering, each character will be distorted. The distortion here is not the rotation of the character but the smooth translation of the upper and lower sides of the character. In the process of rendering, the character will retain a certain spacing according to the situation. When it is negative, it indicates two-character adhesion; this can avoid the problem of poor user experience caused by traditional character rotation, and can ensure that the character recognition of the character has a higher recognition rate.

In step S205, the interference line is generated. The interference line is added as an interference element to the picture generated in step S204, wherein each interference line is a cubic equation track determined by four random points in the picture, and the length of the interference line is related to the width and length of the picture. And because it is determined by four random points, its trajectory will change randomly. The interference line is the same color as the character, and the thickness of the interference line is gradually changed within a certain range, and the width of the line is ensured between the maximum width and the minimum width of the character, which can prevent the removal of the interference line during the verification code cracking, thereby ensuring verification. The difficulty of code cracking.

FIG. 3 is a structural block diagram of a picture verification code generation system according to the present invention, including:

The character string selection module 301 is configured to: randomly select, from the character set, a set of character strings including a plurality of characters, where the length of the character string is randomly determined;

a font rendering module 302, configured to: select each character included in the character string from the word The font set randomly selects a font and a font size for rendering processing to obtain rendered characters, the font library includes a plurality of fonts, and each font has a plurality of font sizes;

a font warping module 303, configured to: perform distortion processing on each of the rendered characters to obtain a twisted character;

The image pre-generation module 304 is configured to: generate a picture including a plurality of the twisted characters;

The interference line adding module 305 is configured to: generate at least one interference line on the picture to obtain a picture verification code, where the trajectory of the interference line is randomly generated.

In one of the embodiments:

And a verification code parameter obtaining module, configured to: in response to the verification code providing request, obtain an application identifier included in the request according to the verification code, and obtain a verification code parameter corresponding to the application identifier, The verification code parameter includes a minimum allowable length and a maximum allowable length of the character string, a picture color, and a picture resolution;

And in the string selection module, the length of the character string is randomly determined between the minimum allowable length and the maximum allowable length;

And in the picture pre-generation module, the picture is determined according to a picture color and a picture resolution included in the verification code parameter.

The above described embodiments only express several embodiments of the present invention, and the description thereof is more The invention is not to be construed as limiting the scope of the invention. It should be noted that a number of variations and modifications may be made by those skilled in the art without departing from the spirit and scope of the invention. Therefore, the scope of the invention should be determined by the appended claims.

Claims

A picture verification code generation method includes:

The string selection step includes: randomly selecting, from the character set, a set of characters including a plurality of characters, the length of the string being randomly determined;

The font rendering step includes: randomly selecting a font and a font from a font set to perform rendering processing on each character included in the character string, the font library includes a plurality of fonts, and each font has multiple types. Font size

The font warping step includes: twisting each of the rendered characters to obtain a twisted character;

a picture pre-generation step, comprising: generating a picture including a plurality of the twisted characters;

The interference line adding step includes: generating at least one interference line on the picture to obtain a picture verification code, and the trajectory of the interference line is randomly generated.
The picture verification code generating method according to claim 1, wherein:

And a verification code parameter obtaining step, the verification code parameter obtaining step includes: obtaining, according to the verification code providing request, the application identifier included in the verification code providing request, and acquiring the verification code parameter corresponding to the application identifier, The verification code parameters include a minimum allowable length and a maximum allowable length of the character string, a picture color, and a picture resolution;

And in the string selection step, the length of the character string is randomly determined between the minimum allowable length and the maximum allowable length;

And in the picture pre-generation step, the picture is determined according to a picture color and a picture resolution included in the verification code parameter.
The picture verification code generating method according to claim 1, wherein the distortion processing comprises: selecting at least one part as a to-be-distorted portion on each of the rendered characters, and performing an upward shift on each of the to-be-distorted portions, Pan down, pan left, or pan right.
The picture verification code generating method according to claim 1, wherein each of said pieces The interference line is generated as follows:

A plurality of track points are randomly selected on the picture to generate a plurality of equation tracks including all of the track points.
The picture verification code generating method according to claim 1, wherein the interference line is the same as the color of the character, and the thickness of the interference line is gradually changed between a maximum width and a minimum width of all the twisted characters.
A picture verification code generation system includes:

a string selection module, configured to: randomly select a group of characters including a plurality of characters from a character set, and the length of the string is randomly determined;

a font rendering module, configured to: randomly select a font and a font size from a font set to render characters for each character included in the character string, where the font library includes multiple fonts, and each font has multiple fonts. Kind size

a font warping module, configured to: perform distortion processing on each of the rendered characters to obtain a twisted character;

a picture pre-generation module, configured to: generate a picture including a plurality of the twisted characters;

The interference line adding module is configured to: generate at least one interference line on the picture to obtain a picture verification code, and the trajectory of the interference line is randomly generated.
The picture verification code generating system according to claim 6, wherein:

And a verification code parameter obtaining module, configured to: in response to the verification code providing request, obtain an application identifier included in the request according to the verification code, and obtain a verification code parameter corresponding to the application identifier, The verification code parameter includes a minimum allowable length and a maximum allowable length of the character string, a picture color, and a picture resolution;

And in the string selection module, the length of the character string is randomly determined between the minimum allowable length and the maximum allowable length;

And in the picture pre-generation module, the picture is determined according to a picture color and a picture resolution included in the verification code parameter.
The picture verification code generating system according to claim 6, wherein the distortion processing comprises: selecting at least one portion as a to-be-distorted portion on each of the rendered characters, and performing an upward shift on each of the to-be-distorted portions, Pan down, pan left, or pan right.
The picture verification code generating system according to claim 6, wherein each of said interference lines is generated in the following manner:

A plurality of track points are randomly selected on the picture to generate a plurality of equation tracks including all of the track points.
The picture verification code generating system according to claim 6, wherein the interference line is the same as the color of the character, and the thickness of the interference line is gradually changed between a maximum width and a minimum width of all the twisted characters.